Research Articles
ASTROBIOLOGY
Volume 10, Number 8, 2010
ª Mary Ann Liebert, Inc.
DOI: 10.1089/ast.2009.9455
The Formation of Sulfate and Elemental Sulfur
Aerosols under Varying Laboratory Conditions:
Implications for Early Earth
H. Langley DeWitt,1,2 Christa A. Hasenkopf,2,3 Melissa G. Trainer,4 Delphine K. Farmer,2
Jose L. Jimenez,1,2 Christopher P. McKay,5 Owen B. Toon,3 and Margaret A. Tolbert1,2
Abstract
The presence of sulfur mass-independent fractionation (S-MIF) in sediments more than 2.45109 years old is
thought to be evidence for an early anoxic atmosphere. Photolysis of sulfur dioxide (SO2) by UV light with
l < 220 nm has been shown in models and some initial laboratory studies to create a S-MIF; however, sulfur must
leave the atmosphere in at least two chemically different forms to preserve any S-MIF signature. Two commonly
cited examples of chemically different sulfur species that could have exited the atmosphere are elemental sulfur
(S8) and sulfuric acid (H2SO4) aerosols. Here, we use real-time aerosol mass spectrometry to directly detect the
sulfur-containing aerosols formed when SO2 either photolyzes at wavelengths from 115 to 400 nm, to simulate
the UV solar spectrum, or interacts with high-energy electrons, to simulate lightning. We found that sulfurcontaining aerosols form under all laboratory conditions. Further, the addition of a reducing gas, in our experiments hydrogen (H2) or methane (CH4), increased the formation of S8. With UV photolysis, formation of S8
aerosols is highly dependent on the initial SO2 pressure; and S8 is only formed at a 2% SO2 mixing ratio and
greater in the absence of a reductant, and at a 0.2% SO2 mixing ratio and greater in the presence of 1000 ppmv
CH4. We also found that organosulfur compounds are formed from the photolysis of CH4 and moderate
amounts of SO2. The implications for sulfur aerosols on early Earth are discussed. Key Words: S-MIF—Archean
atmosphere—Early Earth—Sulfur aerosols. Astrobiology 10, 773–781.
1. Introduction
U
nderstanding the atmosphere of early Earth during
the Archean, the period of time approximately 4–2.45
billion years ago, is an important part of understanding the
conditions under which life originated and developed. Not
much is known with certainty about the atmosphere during
the Archean. However, various pieces of geological evidence
provide guidelines toward the atmospheric composition and
Earth’s climate at the time when life emerged on Earth.
It is thought that a major change in the oxidative state of
early Earth’s atmosphere occurred at the end of the Archean,
approximately 2.45 Ga (Bekker et al., 2004, and references
therein), with perhaps the strongest evidence coming from
isotopic patterns seen in S-bearing sediments through time
(Farquhar et al., 2000). Sulfur is emitted from volcanic ac-
tivity as either hydrogen sulfide (H2S) or sulfur dioxide (SO2)
and readily converts between the two different starting
species in a reducing atmosphere (Kasting et al., 1989).
Around 2 ppbv, SO2 could have existed in the reducing atmosphere of early Earth if the volcanic outgassing rate was
three times as high as it is today, assuming the ocean was
saturated with SO2 and thus rainout was not a loss process
for SO2 (Kasting et al., 1989). Measurements show that sulfur
mass-independent fractionation (S-MIF) occurs in sedimentary rocks older than 2.45 Ga. However, all the S-isotopes fall
along a mass-dependent fractionation line in younger sediments (Farquhar et al., 2000). An anoxic early Earth atmosphere could help explain these data (Pavlov and Kasting,
2002).
Laboratory experiments (Farquhar et al., 2001) and model
simulations (Lyons, 2007) suggest that the photolysis of SO2
1
Department of Chemistry and Biochemistry, University of Colorado at Boulder, Boulder, Colorado.
Cooperative Institute for Research in the Environmental Sciences, University of Colorado at Boulder, Boulder, Colorado.
3
Department of Atmospheric and Oceanic Sciences, Laboratory for Atmospheric and Space Physics, University of Colorado at Boulder,
Boulder, Colorado.
4
NASA Goddard Space Flight Center, Greenbelt, Maryland.
5
Space Science Division, NASA Ames Research Center, Moffett Field, California.
2
773
774
by UV light with a wavelength <220 nm at low altitudes is
one way to form the S-MIF found in the Archean sediments.
For these wavelengths of UV light to reach the lower atmosphere, an anoxic early Earth atmosphere would have been
necessary (Farquhar et al., 2001). The formation of S-MIF
below 10 km is believed to be necessary for the efficient
fallout of these S-MIF–bearing species to reach the ground
(Domagal-Goldman et al., 2008). Ozone (O3) and oxygen (O2)
in the current stratosphere absorb the lower-wavelength incoming solar radiation and prevent the wavelengths of light
necessary to form S-MIF from reaching the lower atmosphere. The preservation of the S-MIF signals in the geological record requires the formation of at least two types of
S-bearing species with different oxidative states (Kasting,
2001). The different sulfur reservoirs must be spatially and
temporally separated to preserve the S-MIF signal (Pavlov
and Kasting, 2002). H2SO4 and elemental sulfur (specifically
S8) aerosols formed from the photolysis and subsequent reactions of SO2 are often cited as two possible reservoirs for SMIF (Kasting et al., 1989; Ono et al., 2003; Zahnle et al., 2006;
Domagal-Goldman et al., 2008). To achieve these two distinct
sulfur reservoirs, it has been proposed that an extremely
anoxic atmosphere is necessary, as the presence of oxygen at
105 times the present atmospheric level (PAL) would
eventually cause all the sulfur in the atmosphere to become
H2SO4 (Pavlov and Kasting, 2002).
The chemical mechanisms used in model predictions of
the relative concentrations of different types of sulfurcontaining aerosols often contain unknown and uncertain rate
constants that may affect the results of the model (Kasting
et al., 1989; Pavlov and Kasting, 2002). Previous experimental
work on S-MIF formation from low-wavelength photolysis
by Farquhar et al. (2001) was performed at high levels of SO2,
approximately 1–2% SO2 in carbon dioxide (CO2) with, and
without, added water (H2O). At these levels of SO2, both S8
and H2SO4 aerosols were formed. However, these levels are
significantly higher than those believed to have been present
on early Earth (Kasting et al., 1989). A model study by Ono
et al. (2003) found that the amount of SO2 greatly influenced
the S8/H2SO4 ratio, with S8 composing about 50% of the total
S-bearing aerosol mass formed at an SO2 outgassing flux of
3.5109 molecules cm2 s1, while almost no S8 was
formed at an SO2 outgassing rate an order of magnitude
lower. Later models have probed the effects of other major
atmospheric constituents on sulfur aerosol chemistry.
Zahnle et al. (2006) used a 1-D model to study the formation
of S8 from photochemical reactions of O2, methane (CH4),
and SO2 in the atmosphere. They found that efficient S8
production at O2 levels of 107 PAL or lower requires the
presence of reduced gases (i.e., CH4 or hydrogen, H2). With
a fixed O2 level of 107 PAL, they found that S8 production
shuts off for CH4 < 8 ppmv. Increasing the CH4 mixing
ratio allowed S8 to form at higher O2 levels (Zahnle et al.,
2006). Significant amounts of CH4 could have come from
early methanogens. Ono et al. (2003) found a similar positive correlation in their model with the atmospheric pressure of CH4 and the formation of S8. Other forms of
S-bearing aerosols, such as organosulfur (a general term for
species containing C and S), could also have been present
under the atmospheric conditions of early Earth and may
have helped preserve S-MIF signal (Lyons, 2009; Ueno et al.,
2009).
DEWITT ET AL.
Here, we report the results of laboratory experiments to
probe the formation of different chemical forms of sulfurcontaining aerosols through UV radiation (to produce aerosols from photolysis from solar radiation) and electrical
discharge (to simulate lightning). We used real-time aerosol
mass spectrometry to probe the chemical composition of
sulfur aerosols formed from a mixture of CH4/CO2/H2 with
SO2. The effect of the initial partial pressure of SO2 was also
investigated.
2. Experimental
2.1. Aerosol generation
The reaction system used to perform these experiments
was discussed in detail previously (Trainer et al., 2004, 2006).
The reactant gases were first combined in a mixing chamber
where they were allowed to equilibrate for 12 hours to ensure homogeneity. The gas mixture was then flowed into a
reaction chamber with a Mykrolis mass flow controller at a
flow rate of 60 standard cubic centimeters per minute. As the
reagent gases passed through the reaction chamber, they
were exposed to either an electrical discharge (Tesla coil) to
simulate lightning or a Hamamatsu broad-spectrum UV
lamp with a wavelength range from 115–400 nm to simulate
the UV region of the solar spectrum. Reactions within each
chamber produced sulfur-containing aerosols. The gas mixtures were only exposed to the narrow line of spark discharge for a few seconds before being flowed through the
reaction cell, similar to a lightning strike. The residence time
for the gas inside the photochemical reaction chamber was 5
minutes. Note that particles in the real atmosphere would
have far longer reaction times than those in the present work.
The mechanical pump on our vacuum can theoretically
pump down to 0.001 torr, which gave us a lower limit of O2
in our system of 106 PAL. This is within O2 levels that
modelers predict should still produce S8, especially with the
addition of reducing gases (Zahnle et al., 2006).
2.2. Particle characterization
Aerosol composition as a function of reagent gas was
measured with a quadrupole-based aerosol mass spectrometer (Q-AMS, Aerodyne Research, Billerica, MA) ( Jayne et al.,
2000; Canagaratna et al., 2007). The reacted particle/gas
mixture flowed from the reaction chamber into the Q-AMS,
where submicron particles were focused into a tight beam
via an aerodynamic lens. The Q-AMS can be used in either
the particle time-of-flight mode or in mass spectrum mode.
The particle time-of-flight mode uses the terminal velocity of
the particles in vacuum to determine their aerodynamic diameter, dva (DeCarlo et al., 2004), while in mass spectrum
mode all particles are analyzed without size selection. The
chemical composition of the particles was determined by
flash vaporizing the particles on a heated tungsten surface at
6008C under high vacuum, followed by electron impact
ionization (70 eV) and mass analysis with a quadrupole mass
spectrometer. With this technique, we were able to measure
the mass spectra of the aerosol compounds in real time
without the need for collection. The Q-AMS was well suited
to these experiments because its inlet and differential
pumping system concentrates the particles by a factor of 107
relative to gaseous compounds, which thus minimizes gas
SULFUR AEROSOLS ON EARLY EARTH
phase interferences. In addition to composition and size information, the Q-AMS also supplies information on the mass
concentration (mg m3) of the particles sampled based on a
calibration of the electron impact ionization signals ( Jimenez
et al., 2003).
For some experiments, a high-resolution time-of-flight
aerosol mass spectrometer (HR-AMS) (DeCarlo et al., 2006)
was also used in conjunction with the standard unit-mass
resolution Q-AMS. The HR-AMS uses the same aerosol inlet,
vaporization, and ionization system as the Q-AMS but with a
time-of-flight mass spectrometer detection system in place of
a quadrupole detector. Two analysis modes can be used for
this system: a V-mode, with an effective ion path length
of 1.3 m; and a W-mode, with an effective ion path length
of 2.9 m. An order of magnitude of sensitivity is gained
by operating in the V-mode, but some resolution is lost
(DeCarlo et al., 2006). In this study, we chose to operate the
HR-AMS solely in the W-mode for maximum peak resolution. While the Q-AMS has a resolving power of m/Dm * 2m
at all mass-to-charge ratios (m/z), with m being the nominal
m/z and Dm being full width at half maximum, the HR-AMS
in W-mode has a maximum resolution of at least m/
Dm ¼ 4300 at 200 m/z, which varies slightly over other m/z
(DeCarlo et al., 2006). This allows identification of peaks that
have the same nominal mass but different chemical formulas.
Thus, more positive identification of fragments arising
from molecules with specific functional groups, such as organosulfur or oxygenated organic molecules, can be made.
2.3. Gas mixtures
Various amounts of trace gases were mixed along with a
nitrogen (N2) background overnight in a mixing chamber.
SO2 (ultra high purity), H2 (ultra high purity), CH4 (research
grade), CO2 (research grade), and N2 (ultra high purity) were
purchased from Airgas Intermountain. A premixed standard
gas of 2000 20 ppmv SO2 in N2 was used as the starting gas
for SO2 in some instances. Pure SO2 was diluted with N2 for
higher mixing ratio experiments with UV photolysis. The
aerosol mass spectrometer (AMS) spectra of each gas mixture
was obtained before initiating reaction and then subtracted
from the aerosol spectra to reduce gas interference in our
results.
The HR-AMS spectrum of the background gases was used
to verify the O2 upper limit. No O2 was detected in the mass
spectrum of the background gas mixture. At our running
conditions, the HR-AMS could detect O2 gas levels of
200 ppmv and above. Thus, an upper limit of O2 in the gas
mixture less than 103 PAL was confirmed with the HRAMS. Combining this calculation with the pumping capability of our vacuum system gives a possible range of 106 to
103 PAL O2 in the reaction chamber.
2.3.1. Electrical discharge. To simulate a volcanic
plume, 20 0.2 ppmv SO2 with, and without, 0.1%
(1000 20 ppmv) CO2 and with varying amounts of H2 were
mixed together. CO2 was used as both an additional oxidant
and a potential carbon source, while H2 was used as an
additional reductant. The overall amount of H2 was varied to
change the average oxidation state of the gas mixtures. CO2
was used in half the experiments. To simplify the aerosol
spectra and their interpretation, we chose H2 instead of CH4
775
as a reductant for the electrical discharge experiments even
though models have shown that both could affect SO2
chemistry in similar ways (Zahnle et al., 2006) and both could
have been present in elevated amounts in the atmosphere of
early Earth (Pavlov et al., 2000; Tian et al., 2005).
2.3.2. Photolysis. We used gas mixtures with varying
amounts of SO2 and CH4 in a N2 background to simulate the
photochemical reactions that potentially occurred in the atmosphere of Archean Earth. We varied the partial pressure
of SO2 from 2 (0.2) to 20,000 (20) ppmv in our experiments to probe the effect of the SO2 mixing ratio on aerosol
formation. SO2 amounts from 2 to 2000 ppmv are not unreasonable for volcanic clouds in early Earth’s atmosphere
following a volcanic eruption. For some experiments, we
added 1000 20 ppmv CH4 to our mixtures, a reasonable
amount of CH4 for the late Archean (Haqq-Misra et al., 2008).
While H2 was added to several gas mixtures, no effect was
observed, likely because H2 does not photolyze at the
wavelengths used in our experiment. Thus, CH4 was used as
the main reducing gas in the photolysis studies.
3. Results
3.1. Electrical discharge
Figure 1a shows the aerosol mass spectra of 20 ppmv SO2
exposed to an electrical discharge. At 20 ppmv SO2, the
majority of the aerosol formed was H2SO4, which resulted
from reaction of SO2 and hydroxyl (OH) radicals formed
from trace H2O. Although the vacuum reaction chamber was
evacuated to under a torr and kept at this level, the trace
amounts of water left in the vacuum chamber or brought
along with the gases heavily influenced the resulting aerosol
composition after the gases were exposed to electrical discharge. Farquhar et al. (2001) also saw H2SO4 in their experiments without added H2O. Increasing H2 from 0 to 10% of the
total gas mixture slowly increased the relative ratio of S8/
H2SO4 in the aerosol formed. As an example, Fig. 1b shows
the mass spectrum with 5% H2. Here, the mass signature for
elemental sulfur is clearly observed. The Q-AMS can be used
to quantify the fraction of S8 formed under varying conditions. Figure 2 shows the relative amounts of characteristic
H2SO4 and S8 peaks from the electrical discharge of 20 ppmv
SO2 as a function of H2 (not including overlapping peaks). It
can be seen that above 4% H2, S8 is the dominant S aerosol
formed. The experiments were repeated with the addition of
1000 20 ppmv (0.1%) CO2 added to the SO2/H2/N2 mixture.
Figure 2 shows that again for this mixture the addition of H2
favors the formation of S8 over H2SO4. However, with the
additional CO2 oxidant added, more H2 is needed to switch
the aerosol formed from mainly H2SO4 to mainly S8.
3.2. UV experiments
3.2.1. The effect of the SO2 mixing ratio on aerosol
chemistry. For the UV experiments, gas mixtures containing varying mixing ratios of SO2 were exposed to a broadspectrum UV light source with wavelengths that ranged
from 115 to 400 nm. Varying amounts of CH4 and H2 were
added to the gas mixtures. The empty reaction chamber was
exposed to UV light for 15 minutes while under vacuum
prior to each experiment to photolyze H2O contamination on
776
DEWITT ET AL.
FIG. 1. Q-AMS spectra of particles
formed from the electrical discharge
of 20 ppmv SO2 in N2, without (a) and
with (b) 50,000 ppmv (5%) H2. Plotted
is fraction of total aerosol mass versus
m/z (amu). Likely peak identifications
have been made. Spectra to the right
of the dashed line have been multiplied by 10 for ease of viewing. Small
signals not from S8 or H2SO4 aerosols
(unlabeled peaks) are around the
noise level of the instrument and
could be from trace contamination
from the gas or flow system.
the walls of the reaction vessel. The reaction cell was also
cooled both before and during reaction to 273 K with a
methanol cooler in an attempt to reduce H2O contamination.
Although we achieved some reduction in H2O, not all the
H2O contamination was removed.
The mixing ratio of SO2 was varied to investigate the effect
of initial partial pressure on the aerosol chemistry. With 2, 20,
200, and 2000 ppmv SO2 in N2, only the H2SO4 signal was
observed. Figure 3a shows a representative spectrum using
2000 ppmv SO2. When SO2 was further increased to 2% of
the total gas mixture, a complete spectral signature for elemental sulfur through S8 was observed in addition to H2SO4
(Fig. 3b). Note that Farquhar et al. (2001) also used 2% SO2 in
N2 and observed the formation of both S8 and H2SO4 aerosols. When SO2 was increased to 4% of the total gas mixture,
a further increase in the elemental sulfur signature was observed (not shown). Two percent SO2 photolyzed at room
temperature rather than 273 K did not form S8, possibly indicating the importance of the increased H2O in the cell at
room temperature.
3.2.2. The effect of reducing gases on SO2 aerosol
chemistry. In addition to exploring the effects of the SO2
mixing ratio on UV photolysis aerosol chemistry, we also
probed the effect of adding reducing gases. Initially 1% H2
was added to a 2000 ppmv SO2 gas mixture, but no change
was observed in the mass spectrum. Although H2 does not
photolyze at any of the wavelengths emitted by the UV
lamp, it is possible that reactions such as H2 þ O ? OH þ H
or H2 þ OH ? H2O þ H could have occurred; thus H2 could
have affected the S-aerosol chemistry (Moses et al., 2000).
However, these reactions are slow, and our experimental
results do not reflect them occurring. Under our experimental
conditions, it is likely that the much faster reaction of SO2 þ
OH þ M ? HSO3 outcompetes the reaction of OH with H2.
In contrast to adding H2 to the initial gas mixture, the
combination of 1000 ppmv CH4 with 2000 ppmv SO2 results
in a strong S8 signal in addition to the H2SO4 signal (Fig. 3c),
with mass signal in the aerosol mass spectrum for S1–S8.
Similar experiments were performed on 2% and 4% SO2 gas
mixtures, both with 1000 ppmv CH4. In these cases, adding
CH4 to the initial gas mixture increased the amount of S8
compared to gas mixtures with the same mixing ratio of SO2
but without CH4. Figure 4 shows a summary of the fraction
of S8 aerosol found as a function of SO2 pressure, both with and
without added CH4. It can be seen that both increasing the
initial SO2 mixing ratio and adding CH4 increase the S8 fraction.
3.2.3. Formation of organosulfur from CH4/SO2 photolysis. While Fig. 3c shows CH4 addition helps the formation
of S8, other S-bearing peaks were observed in the AMS
spectrum, especially at lower levels of SO2. Figure 5 shows
the resulting spectrum from the photolysis of 2 ppmv SO2
and 1000 ppmv CH4. Although the prevalent mass peaks for
the 2 ppmv SO2 mixture followed a H2SO4 and a separate
organic ion series pattern, organosulfur species have been
FIG. 2. Relative fraction of total S-bearing aerosol signal of
characteristic H2SO4 peaks (crosses) and S8 peaks (circles)
versus percent H2 calculated from Q-AMS data. Experiments
with CO2 are shown with a dashed line. Overlapping mass
peaks for H2SO4 and S8 (m/z 32 and 64) have not been included in these calculations.
SULFUR AEROSOLS ON EARLY EARTH
777
FIG. 3. Q-AMS spectra of particles formed from the photolysis of 2000 ppmv (0.2%) SO2 in N2 (a); 20,000 ppmv SO2 (b); and
2000 ppmv SO2 with 1000 ppmv CH4 (c). Plotted is fraction of total aerosol mass versus m/z (amu). Likely peak identification
has been made. Peaks to the right of the dashed line have been multiplied by 100 for ease of viewing. Small signals not from
S8 or H2SO4 aerosols (unlabeled peaks) for mass spectra without CH4 are around the noise level of the instrument and could
be from trace contamination from the gas or flow system.
shown to fragment in the HR-AMS to predominately hydrocarbon and H2SO4 peaks. No S8 was observed at this low
level of SO2. However, some amount of organosulfur was
observed in the HR-AMS data. Specifically, CH3 SO2þ (HRAMS spectra shown in Fig. 5 inset), CH3SOþ, and CH3Sþ
were observed. We also observed HSþ and H2Sþ peaks. The
CH3 SO2þ peak is prominent in the AMS spectrum of methyl
sulfonic acid (MSA, CH3SO3H), and the fragmentation of
MSA in the AMS is well characterized (Zorn et al., 2008).
Using MSA as a model for organosulfur species, one can
determine a CH3 SO2þ =H2 SO4 ratio of 0.25 and apply this
ratio to our data (Zorn et al., 2008). This allows one to
roughly quantify the amount of sulfate signal due to an organosulfur compound such as MSA as opposed to that from
inorganic sulfate. This is illustrated in Fig. 5a. From this
calculation, it can be seen that approximately 16% of the
sulfate signal could be due to MSA fragmentation for 2 ppmv
SO2 with 1000 ppmv CH4.
The spectrum was also compared to a standard organosulfate spectrum (trihydroxy sulfate ester, C5H12O7S) by
normalizing the organosulfate spectrum’s CH3 SO2þ peak to
that in the spectrum of 2 ppmv SO2 with 1000 ppmv CH4
(Farmer et al., 2010). When comparing the resulting spectra,
the normalized standard organosulfate spectrum’s sulfate
signal is approximately 7 times larger than that of the spectrum from these experiments. Thus, 14% of our CH3 SO2þ peak
could be due to organosulfate, with the rest of the signal
from other less-oxidized organosulfur species such as MSA.
Although the spectrum for the 20 ppmv SO2 mixture
looked very similar to that of the 2 ppmv SO2 mixture, there
was an approximate factor of 2 reduction in the relative
prominence of organosulfur peaks as compared to H2SO4
peaks with increased SO2.
Note that no S8 was observed from the photolysis of
1000 ppmv CH4 with 20 ppmv and below SO2. All the UV
experiments and results have been summarized in Table 1.
4. Discussion
Our studies allowed real-time analysis of nonrefractory
aerosol compound produced in the flow system simulating
the aerosol potentially formed from volcanic emissions in an
early Earth atmosphere. In these experiments, trace amounts
778
FIG. 4. Percent of S-bearing aerosol mass from S8 as a
function of the pressure of SO2 calculated from Q-AMS data.
Experiments are shown with (circles) and without (triangles)
1000 ppmv CH4. Overlapping mass peaks for H2SO4 and S8
(m/z 32 and 64) have not been included in these calculations.
of H2O in the vacuum reaction cell were enough to oxidize
most of the sulfur to H2SO4. We cooled our reaction vessel to
273 K. This gives us a theoretical water vapor concentration
of 0.7%, or 7000 ppmv within the reaction cell as an upper
limit. The HR-AMS background gas spectrum further limits
the H2O contamination to 1000 ppmv. For lower altitudes
DEWITT ET AL.
and within volcanic plumes this would be relevant, as water
would be present in at least trace amounts at low altitudes
and would also be emitted along with SO2 from volcanoes
(Pavlov et al., 2000; Textor et al., 2003). At 2% SO2, photolysis
in a room-temperature reaction cell did not yield S8, illustrating that the water vapor in the reaction cell was affecting
the aerosol chemistry.
In our discharge experiments, we studied a simple system
in which we increased the reducing gases, in this case H2, to
avoid the complication of organic chemistry from CH4. The
addition of H2 to the gas mixture caused elemental sulfur
aerosol to form in addition to H2SO4. Increasing the amount
of H2 increased the S8/H2SO4 ratio. Zahnle et al. (2006) found
that increasing the reduced gases in his 1-D model of early
Earth’s atmosphere increased the S8/H2SO4 ratio, and our
results appear to support this finding. One explanation for
this is that H formed from the electrical dissociation of H2
titrated the OH formed from H2O. H2 could also titrate the O
atoms from SO2 photolysis, facilitating the formation of more
reduced sulfur compounds. Alternatively, intermediate HSn
compounds may increase S chain formation (Kasting et al.,
1989). Excess H2 would contribute to the formation of HSn
compounds and partially explain our results. The same effect, albeit dampened, is observed when CO2 is added to the
gas mixture. The increased O from CO2 photolysis most
likely explains the lower amounts of S8 formed with CO2
than without for the same level of H2. Our experiments do
not allow us to determine the isotopic fractionation of SO2,
and in fact electrical discharge may not even produce S-MIF.
FIG. 5. AMS spectrum of particles formed from the photolysis of 2 ppmv SO2 with 1000 ppmv CH4 in N2 (gray boxes) with
the standard AMS spectrum of MSA (black lines). The two spectra are normalized so that the intensity of CH3 SO2þ = (m=z 79)
is equal in each spectrum. Patterned boxes are the sulfate contribution from MSA. Approximately 16% of the sulfate signal in
the 2 ppmv SO2 1000 ppmv CH4 spectra could be from MSA. The expanded high-resolution peak from the organosulfur peak
CH3 SO2þ is shown in the inset.
SULFUR AEROSOLS ON EARLY EARTH
779
Table 1. Summary of SO2 UV Photolysis Experiments
Experimental
SO2 (ppmv)
Reducing gas?
S8?
2 0.02
1,000 20 ppmv CH4
No
20 0.2
1,000 ppmv CH4
200 2
200 2
2,000 20
2,000 20
2,000 20
20,000 20
20,000 20
40,000 20
40,000 20
Amount present
on early Earth
No
1,000 ppmv CH4
No
10,000 20 ppmv H2
1,000 ppmv CH4
No
1,000 ppmv CH4
No
1,000 ppmv CH4
SO2: 10 ppbva
(global lower limit)
Organosulfur?
Yes, present
and studied
No
Yes, present
and studied
No
Not studied
No
Not studied
No
Not studied
No
Not studied
Yes
Not studied
Yes
Not studied
Yes
Not studied
Yes
Not studied
Yes
Not studied
O2: modeled to
H2O: Unknown,
likely >7,000 ppmv
be from 105 ppmvc
b
at low altitudes
to 1013 ppmvb
H2O level
in experiment
(upper limit)
O2 level in
experiment, PAL
(upper limit)
Unknown
103
7,000 ppmv
103
7,000 ppmv
7,000 ppmv
7,000 ppmv
7,000 ppmv
7,000 ppmv
7,000 ppmv
7,000 ppmv
7,000 ppmv
7,000 ppmv
CH4: up to
1,000 ppmvd
103
103
103
103
103
103
103
103
103
a
Kasting et al. (1989).
Pavlov and Kasting (2002).
c
Farquhar et al. (2000).
d
Haqq-Misra et al. (2008).
b
We observe the same effect in our photolysis experiments
when we add CH4 to the initial gas mixtures. With the addition of 1000 ppmv CH4, S8 is formed at 2000 ppmv initial
SO2 mixing ratio. Adding 1000 ppmv CH4 to 2% and 4% SO2
gas mixtures increases S8 formation by a factor of 2 or more.
We infer from these experiments that the relative oxidative
capacity, or amount of reducing gases relative to oxidizing
gases, in a volcanic plume and the ambient atmosphere can
greatly change the chemical composition of the aerosols
formed in and around a volcanic cloud.
The experimental paper by Farquhar et al. (2001) showed
the formation of S8 and H2SO4 aerosols with UV photolysis
but with unrealistically high concentrations of SO2 and noncontinuum UV light sources. In our experiments, we exposed
SO2, CH4, and CO2 to a broad-spectrum UV light with a
peak in emission at *145 nm, under the 220 nm cutoff determined by Farquhar et al. (2001) to be the wavelength of
UV for S-MIF formation. We probed the effect that the initial
SO2 pressure had on the resulting sulfur aerosol chemistry.
We observed S8 formation with 2% SO2 but only H2SO4
aerosol formation with 200 ppmv and lower SO2. While the
lack of S8 production at lower SO2 values may reflect the
importance of the SO2 self-reaction, it may partially also reflect the relative abundance of SO2 and water vapor in the
system. Models predict efficient S8 formation at much lower
levels of SO2 than we observe experimentally, at levels of
10 ppbv (Kasting et al., 1989) in models versus 2000 ppmv or
higher in our experiments. Our experimental photochemical
reaction time of 5 minutes is significantly lower than reaction times in the atmosphere, and this could also have affected our results. The rate constants are not known well
enough to determine how much of an effect reaction time
would have on S8 production (Kasting et al., 1989). Although
our theoretical lower limit for O2 in our vacuum chamber
was on the order of 106 PAL, we were only able to experimentally determine a 103 PAL O2 upper limit after our
experiments began, giving us a wide range of possible O2
levels, which could have also affected our results. Ono et al.
(2003) also predicted a SO2 pressure dependence on the S8/
H2SO4 ratio, which they attributed to the increased S atom
requirement to form S8 over H2SO4. In our system, S þ þ S
must combine to form S2, S2 þ S2 to form S4, and so on. These
reactions are also sensitive to total pressure, as the combination reactions of S atoms to form S chains are three-body
reactions (Kasting et al., 1989). However, the possible pressure dependence of sulfur aerosol formation was not probed
FIG. 6. A conceptual model of the formation and preservation of S-MIF in early Earth’s atmosphere under different
atmospheric conditions. More H2O increases the formation
of H2SO4, while moderate amounts of CH4 form more organosulfur and S8 aerosols. A high CH4/SO2 ratio increases
the amount of organosulfur formed. Increasing the mixing
ratio of SO2 increases the amount of S8 formed, as does
adding CH4, once the SO2 pressure is at a certain threshold.
780
in the present work, as all our experiments were conducted
at 600 torr.
Models have suggested that O2 levels of 105 PAL or greater
will shut off S-MIF formation (Pavlov and Kasting, 2002).
Below this level, O2 would not be well mixed (Pavlov and
Kasting, 2002). We calculate that our mixing chamber has an
upper limit of O2 at 103 PAL. We still observe the formation
of elemental sulfur aerosols under certain experimental conditions with both electrical discharge and photolysis as the
driving force for these reactions. One reason for this could be
that our initial SO2 levels are significantly higher than the atmospheric levels in the models used to predict O2 levels. For
example, we observed the formation of S8 when 2000 ppmv
SO2 was used for the photolysis experiments, a much higher
amount of SO2 than the 10 ppbv SO2 that is predicted to have
been present in early Earth’s atmosphere. In addition, we
added reducing gases such as H2 or CH4.
Organosulfur formed from the photolysis of the biologically produced gases CS2 and OCS have been found to
contain S-MIF (Lyons, 2009), and our research points to an
additional abiological source of organosulfur aerosols in
early Earth’s atmosphere. Ueno et al. (2009) attributed the
buildup of OCS in the atmosphere to the reaction
CO þ Sn þ M ? OCS. In this research, we have found that it
can be difficult to build up significant amounts of Sn. However, we found organosulfur to be present in aerosols formed
from the photolysis of moderate amounts of SO2 and CH4.
Although the signal from organosulfur peaks was relatively
small compared to the sulfate signal in our spectra, organosulfur species have been shown to fragment in the AMS to
peaks typically associated separately with organics and
H2SO4. Using MSA as a model for organosulfur aerosols, we
find that at least 16% of our sulfate aerosol signal could come
from organosulfur for a gas mixture of 2 ppmv SO2 and
1000 ppmv CH4. Organosulfate species fractionate even more
strongly into sulfate and organic peaks in the AMS, with a
greater sulfate/organosulfur ratio than we observe in our
laboratory spectrum. By comparing a standard organosulfate
spectrum to our spectrum, we calculated an upper limit of 14%
of our organosulfur signal could be from organosulfate. Additionally, we find evidence of CH3Sþ peaks in our HR-AMS
spectrum, which indicates that some of our organosulfur signal must come from species with a C–S bond, such as sulfonic
acids or thiols (H–S–R). Organosulfate has only C–O–S bonds.
Observing SHþ ions in our spectrum may also suggest that we
have some signal contribution from thiols. Thus, a substantial
percentage of our AMS sulfate signal is likely due to the fragmentation of organosulfur compounds. Organosulfur aerosols
could have acted as another reservoir for S-MIF and may require less initial SO2 pressure, with possibly less restriction of
O2 levels in early Earth’s atmosphere than the formation of S8.
DEWITT ET AL.
From our experiments, it appears that either high levels of
SO2 or the presence of reducing gases are necessary for the
formation of S8 in the presence of water vapor and lowwavelength UV light. The water vapor levels in our experimental chamber are most likely lower than, or close to, the
levels at lower altitudes in the Archean atmosphere, and
levels of OH in our chamber are probably directly related to
the amount of water vapor. However, models calculate that
OH levels in the Archean atmosphere would have been
limited by reaction with H2, CO, and CH4 (Pavlov et al.,
2000). The increase in S8 with the addition of reducing gases
in our experiments appears to agree with these model results. It is important to note that we did form S8 at levels of
O2 > 105 PAL with sufficiently large mixing ratios of SO2 or
sufficient amounts of reducing gases, or both. An atmosphere with more CH4 or H2 could potentially form more S8
than one that is less reducing. The photolysis products of SO2
and CH4 can also react to form substantial amounts of organosulfur compounds. Lower mixing ratios of SO2 are required for the formation of organosulfur aerosols than would
be required for S8 aerosols; 2 ppmv SO2 or below are required for organosulfur formation, while 2000 ppmv SO2 or
above are needed for S8 aerosol formation. Thus, organosulfur aerosols could provide another reservoir for S-MIF at
various oxidation levels different than S8 and H2SO4, and in
addition can form at more reasonable levels of SO2 in the
presence of H2O. These experiments illustrate that the atmospheric sulfur and methane cycles may have been coupled; thus important chemistry may be missing from
present models of the atmosphere of early Earth. Further
work should be carried out to explore the total pressure
dependence and water vapor dependence of the reactions
that lead to the production of S8, H2SO4, and organosulfur
aerosols.
Acknowledgments
H. Langley DeWitt was partially supported by a NASA
GSRP fellowship. Christa A. Hasenkopf was supported with
a National Science Foundation Graduate Research Fellowship.
We gratefully acknowledge NASA Grant NNX07AV55G for
funding. We thank Donna Sueper (CU-Boulder) for the data
analysis software for the high-resolution AMS.
Abbreviations
AMS, aerosol mass spectrometer; HR-AMS, highresolution time-of-flight aerosol mass spectrometer; MSA,
methyl sulfonic acid; m/z, mass-to-charge ratio; PAL, present
atmospheric level; Q-AMS, quadrupole-based aerosol mass
spectrometer; S-MIF, sulfur mass-independent fractionation.
References
5. Conclusions
Our experiments found that the formation of S8, H2SO4,
and organosulfur aerosols from SO2 is affected by several
experimental factors: (1) initial partial pressure of SO2, (2)
reducing gases present during the reaction of SO2, (3) the
energy source used to initiate the reaction, and (4) the
amount of water vapor present. We have summarized our
conclusions in Fig. 6, modifying a figure by Ono et al. (2003)
to reflect our results.
Bekker, A., Holland, H.D., Wang, P.L., Rumble, D., Stein, H.J.,
Hannah, J.L., Coetzee, L.L., and Beukes, N.J. (2004) Dating the
rise of atmospheric oxygen. Nature 427:117–120.
Canagaratna, M.R., Jayne, J.T., Jimenez, J.L., Allan, J.D., Alfarra,
M.R., Zhang, Q., Onasch, T.B., Drewnick, F., Coe, H., Middlebrook, A., Delia, A., Williams, L.R., Trimborn, A.M.,
Northway, M.J., DeCarlo, P.F., Kolb, C.E., Davidovits, P., and
Worsnop, D.R. (2007) Chemical and microphysical characterization of ambient aerosols with the Aerodyne aerosol mass
spectrometer. Mass Spectrom Rev 26:185–222.
SULFUR AEROSOLS ON EARLY EARTH
DeCarlo, P.F., Slowik, J.G., Worsnop, D.R., Davidovits, P., and
Jimenez, J.L. (2004) Particle morphology and density characterization by combined mobility and aerodynamic diameter
measurements. Part 1: theory. Aerosol Sci Technol 38:1185–
1205.
DeCarlo, P.F., Kimmel, J.R., Trimborn, A., Northway, M.J.,
Jayne, J.T., Aiken, A.C., Gonin, M., Fuhrer, K., Horvath, T.,
Docherty, K.S., Worsnop, D.R., and Jimenez, J.L. (2006) Fielddeployable, high-resolution, time-of-flight aerosol mass spectrometer. Anal Chem 78:8281–8289.
Domagal-Goldman, S.D., Kasting, J.F., Johnston, D.T., and Farquhar, J. (2008) Organic haze, glaciations and multiple sulfur
isotopes in the Mid-Archean Era. Earth Planet Sci Lett 269:29–40.
Farmer, D.K., Matsunaga, A., Docherty K.S., Surratt, J.D., Seinfeld, J.H., Ziemann, P.J., and Jimenez, J.L. (2010) Response of
an aerosol mass spectrometer to organonitrates and organosulfates and implications for atmospheric chemistry. Proc Natl
Acad Sci USA 107:6670–6675.
Farquhar, J., Bao, H.M., and Thiemens, M. (2000) Atmospheric
influence of Earth’s earliest sulfur cycle. Science 289:756–758.
Farquhar, J., Savarino, J., Airieau, S., and Thiemens, M.H. (2001)
Observation of wavelength-sensitive mass-independent sulfur
isotope effects during SO2 photolysis: implications for the
early atmosphere. J Geophys Res 106:32829–32839.
Haqq-Misra, J.D., Domagal-Goldman, S.D., Kasting, P.J., and
Kasting, J.F. (2008) A revised, hazy methane greenhouse for
the Archean Earth. Astrobiology 8:1127–1137.
Jayne, J.T., Leard, D.C., Zhang, X.F., Davidovits, P., Smith, K.A.,
Kolb, C.E., and Worsnop, D.R. (2000) Development of an
aerosol mass spectrometer for size and composition analysis
of submicron particles. Aerosol Sci Technol 33:49–70.
Jimenez, J.L., Jayne, J.T., Shi, Q., Kolb, C.E., Worsnop, D.R.,
Yourshaw, I., Seinfeld, J.H., Flagan, R.C., Zhang, X.F., Smith,
K.A., Morris, J.W., and Davidovits, P. (2003) Ambient aerosol
sampling using the Aerodyne Aerosol Mass Spectrometer. J
Geophys Res 108, doi:10.1029/2001JD001213.
Kasting, J.F. (2001) Earth history—the rise of atmospheric oxygen. Science 293:819–820.
Kasting, J.F., Zahnle, K.J., Pinto, J.P., and Young, A.T. (1989)
Sulfur, ultraviolet-radiation, and the early evolution of life.
Orig Life Evol Biosph 19:95–108.
Lyons, J.R. (2007) Mass-independent fractionation of sulfur isotopes by isotope-selective photodissociation of SO2. Geophys
Res Lett 34, doi:10.1029/2007GL031031.
Lyons, J.R. (2009) Evaluating SO2 photolysis as the source of
Archean sulfur MIF. Geochim Cosmochim Acta 73:A807.
Moses, J.L., Lellouch, E., Bezard, B., Gladstone, G.R., Feuchtgruber, H., and Allen, M. (2000) Photochemistry of Saturn’s
atmosphere: II. Effects of an influx of external oxygen. Icarus
145:166–202.
781
Ono, S., Eigenbrode, J.L., Pavlov, A.A., Kharecha, P., Rumble,
D., Kasting, J.F., and Freeman, K.H. (2003) New insights into
Archean sulfur cycle from mass-independent sulfur isotope
records from the Hamersley Basin, Australia. Earth Planet Sci
Lett 213:15–30.
Pavlov, A.A. and Kasting, J.F. (2002) Mass-independent fractionation of sulfur isotopes in Archean sediments: strong evidence for an anoxic Archean atmosphere. Astrobiology 2:27–41.
Pavlov, A.A., Kasting, J.F., Brown, L.L., Rages, K.A., and
Freedman, R. (2000) Greenhouse warming by CH4 in the atmosphere of early Earth. J Geophys Res 105:11981–11990.
Textor, C., Graf, H.-F., Herzog, M., and Oberhuber, J.M. (2003)
Injection of gases into the stratosphere by explosive volcanic
eruptions. J Geophys Res 108, doi:10.1029/2002JD002987.
Tian, F., Toon, O.B., Pavolv, A.A., and De Sterck, H. (2005) A
hydrogen rich early Earth atmosphere. Science 308:1014–
1017.
Trainer, M.G., Pavlov, A.A., Curtis, D.B., McKay, C.P., Worsnop,
D.R., Delia, A.E., Toohey, D.W., Toon, O.B., and Tolbert, M.A.
(2004) Haze aerosols in the atmosphere of early Earth: manna
from heaven. Abstr Pap Am Chem Soc S 228:U696.
Trainer, M.G., Pavlov, A.A., DeWitt, H.L., Jimenez, J.L., McKay,
C.P., Toon, O.B., and Tolbert, M.A. (2006) Organic haze on
Titan and the early Earth. Proc Natl Acad Sci USA 103:18035–
18042.
Ueno, Y., Johnson, M.S., Danielache, S.O., Eskebjerg, C., Pandey,
A., and Yoshida, N. (2009) Geological sulfur isotopes indicate elevated OCS in the Archean atmosphere, solving
faint young Sun paradox. Proc Natl Acad Sci USA 106:14784–
14789.
Zahnle, K., Claire, M., and Catling, D. (2006) The loss of massindependent fractionation in sulfur due to a Palaeoproterozoic
collapse of atmospheric methane. Geobiology 4:271–283.
Zorn, S.R., Drewnick, F., Schott, M., Hoffmann, T., and Borrmann, S. (2008) Characterization of the South Atlantic marine
boundary layer aerosol using an Aerodyne aerosol mass
spectrometer. Atmospheric Chemistry and Physics 8:4711–4728.
Address correspondence to:
H. Langley DeWitt
CIRES
UCB 216
University of Colorado at Boulder
Boulder, CO 80309
E-mail: [email protected]
Submitted 12 August 2009
Accepted 25 May 2010